B REDD+ BE -REDD I

B

REDD+

DAUWE, T., CUYPERS, D., MEYFROIDT, P., SWINNEN, E., VANGOIDSENHOVEN,M.,VERBIST,B.

Climate

Promotors:

Dauwe Tom (VITO) Meyfroidt Patrick (UCL)

Nyssen Jan (UGent) Swinnen Else (VITO) Verbist Bruno (KUL)

Authors:

Dauwe, T.¹, Cuypers, D.¹, Meyfroidt, P.², Nyssen, J.³, Swinnen, E.⁴, Vangoidsenhoven, M.¹, Verbist, B.⁵

1. VITO, unit Transition, Energy and Environment, Boeretang 200, 2400 Mol; 2. UCL, Earth and Life Institute, Chemin du Cyclotron 2, 1348 Louvain-la-Neuve; 3. UGent, Department of Geography, Krijgslaan 281, 9000 Gent; 4. KUL, Division of Forest, Nature and Landscape, Celestijnenlaan 200, 3001 Heverlee; 5. VITO, unit Remote Sensing, Boeretang 200, 2400 Mol, Belgium.

(SSD)

FINAL REPORT

“

-REDD

”

B

REDD+

Belgian Science Policy Office Avenue Louise 231

Louizalaan 231 B-1050 Brussels Belgium

Tel: +32 (0)2 238 34 11 – Fax: +32 (0)2 230 59 12 http://www.belspo.be

Contact person: Aline Van der Werf +32 (0)2 238 36 71

Neither the Belgian Science Policy nor any person acting on behalf of the Belgian Science Policy is responsible for the use which might be made of the following information. The authors are responsible for the content.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without indicating the reference :

Dauwe, T., Cuypers, D., Meyfroidt, P., Swinnen, E., Vangoidsenhoven, M., Verbist, B., Be-REDDi. Final Report. Brussels: Belgian Science Policy, 2012 – 74 p. (Research Programme Science for a Sustainable Development)

T able of content

Summary ...

Table of content ... 4

List of abbreviations ... 6

Introduction ... 7

1. REDD+ in the UNFCCC ... 9

1.1. Policy options for REDD+ _____________________________________________________ 9 Introduction ... 9

The scope of REDD ... 10

Accounting Basis ... 12

Monitoring, Reporting and Verification Scheme ... 15

Financing Scheme ... 16

Scale... 19

1.2. REDD+ negotiations: a state-of-play ___________________________________________ 22 From Montreal to Copenhagen ... 22

The Cancun agreement ... 23

COP17 in Durban: progress on how to measure REDD+ ... 27

2. How can REDD+ be sustainable ... 30

2.1. Imbedding REDD+ in the forest REALU’ty _______________________________________ 30 Forest definition and leakage ... 30

Biodiversity ... 30

Ecosystem carbon and REALU: an alternative approach ... 31

Mitigation … ... 31

… and poverty reduction by adaptation ... 32

Monitoring-Reporting-Verification ... 32

3. Potential of low resolution optical remote sensing imagery ... 34

Introduction ... 34

Net Primary Productivity (NPP) ... 36

Light Use Efficiency ... 39

4. Drivers of deforestation and land use change, and policies for REDD+ and global forest transition ... 44

Introduction ... 44

Current state and future trends of global land use ... 44

State of the knowledge on forest transitions ... 46

Policies for REDD+ and Forest transitions ... 49

5. The opportunity cost of REDD+ ... 54

What are the costs of REDD+? ____________________________________________________ 54

RED, REDD and REDD+ ... 55

The eye of the beholder – the accounting stance ... 55

Global cost estimates ... 55

The implementation cost ________________________________________________________ 56 The transaction costs ___________________________________________________________ 57 Stabilization costs ______________________________________________________________ 57 Opportunity costs ______________________________________________________________ 58 Indirect opportunity costs ... 58

The demand curve ... 59

The discount rate ... 59

Opportunity cost estimates ... 59

The difference between cost efficient and easy _______________________________________ 60 References ... 62

Annex ... 71

Dissemination _________________________________________________________________ 71 Working papers ... 71

Policy Briefs ... 71

Research papers ... 71

Conferences attended ... 72

Conference organised ... 73 Additional funding ______________________________________________________________ 74

L ist of abbreviations

A/R Afforestation/Reforestation AAU Assigned Amount Units

AFOLU Agriculture, forestry and land use BAP Bali Action Plan

BAU Business-as-usual

CDM Clean Development Mechanism CER Certified Emission Reduction CfRN Coalition for Rainforest Nations COP Conference of the Parties

EC Eddy Covariance

ERU Emission Reduction Unit ETS EU Emission Trading Scheme

EU European Union

FCPF Forest Carbon Partnership Facility GPG Good Practice Guidelines

GPP Gross Primary Productivity

IPCC Intergovernmental Panel on Climate Change LUCC Land Use and Land Cover

LULUCF Land Use, Land Use Change and Forestry MRV Monitoring, Reporting and Verification NAMA National Appropriate Mitigation Action NBP Net Biome Productivity

NEE Net Ecosystem Exchange NEP Net Ecosystem Productivity NGO Non-Governmental Organisation NPP Net Primary Productivity

PAM Policies and measures

REALU Reducing Emissions from All Land Uses

REDD Reducing Emissions from Deforestation and forest Degradation in developing countries

TDERU Tropical Deforestation Emission Reduction Unit

UNFCCC United Nations Framework Convention on Climate Change

I ntroduction

The United Nations Framework Convention on Climate Change (UNFCCC) was established to avoid negative impacts of climate change through prevention, mitigation and adaptation. Deforestation and degradation of forests in developing countries contributes to approximately 12 to 17% of the global annual greenhouse gas emissions (IPCC, 2007; van der Werf et al., 2009). With such a large share, reducing emissions in the Land-Use, Land Use Change and Forestry (LULUCF) sector is essential in climate change mitigation. Despite of being a substantial opportunity for achieving significant greenhouse gas emission reductions, avoided deforestation projects were excluded from the Kyoto Protocol’s Clean Development Mechanism (CDM), because of methodological and political reasons (Fry, 2008). Furthermore, there were concerns that including a project-based mechanism that could avoid deforestation in the CDM could result in significant leakage, permanence problems, flooding the market with cheap credits, massive offsetting through this mechanism and additionality problems (Fry, 2008). Thus, for the first commitment period, LULUCF mitigation options in developing countries were limited to CDM afforestation and reforestation projects (Schlamadinger et al., 2007).

Avoiding deforestation was however not of the negotiating table and during the UNFCCC meeting in Montreal in 2005, the Coalition for Rainforest Nations (CfRN) proposed a new mechanism to deal with emissions from deforestation. Since then discussions on Reduced Emissions from Deforestation and forest Degradation in Developing Countries (REDD) have progressed considerably. The objective of a REDD payment distribution mechanisms is to support policies and measures that reduce deforestation and degradation through the transfer of revenues from international REDD funds or carbon markets to (or within) national levels. This may provide benefits of three types:

a) shared responsibility for reducing a major driver of global climate change,

b) financial payments and co-investment that exceed the economic opportunities foregone from decisions to maintain carbon stocks, and

c) co-benefits through the other environmental service functions that well-maintained forests can provide.

To ensure demonstrable results on emission reduction, REDD must be effective in targeting the wide range of agents involved in deforestation and degradation (drivers). Therefore they must incentivize and reward good performances compared to reference scenarios and adequately compensate agents that suffer losses from changed practices.

Ultimately, the objective of a REDD scheme is to scientifically, technologically and financially support developing countries to install policies and measures that reduce deforestation and degradation. To ensure verifiable emission reductions, effective forest monitoring systems must be installed in participating developing countries. These systems will have to combine remote sensing imagery and field data. A particular difficulty for developing countries will be to establish an effective and cost- efficient system to monitor and assess the progress that has been made in the different activities that are eligible under REDD.

In the end, REDD can only be successful if developing countries are able to deal with the agents that are driving deforestation and forest degradation. REDD can thus only be viewed in the larger context of agriculture, forestry and land use (AFOLU). Developing countries have the extremely difficult task to protect and conserve their forests in an age when demand for forest products and agricultural commodities will only increase due to increasing population (especially in developing countries), increased welfare (and subsequent consumption) and in some regions declined agricultural productivity due to climate changes and other environmental problems (e.g. soil erosion). Part of the

solution could be to move from REDD to REALU (reducing emissions from all land uses), abandoning the sometimes artificial distinction between forest and non-forest. REALU would give developing countries an incentive to look at land use in an holistic perspective.

Any REDD scheme will however come with a considerable cost to developing countries. Not only costs to national or sub-national governments in implementing policies and measures (PAM) and setting up an effective forest monitoring system, but also the foregone economic benefits from further deforestation (opportunity costs). An international REDD scheme that is not able to attract sufficient funds from developed countries or international carbon markets to provide an incentive to all developing countries to keep forests intact, is deemed to be ineffective as it will lead to leakage.

In this report, a comprehensive review is given of the different options for the REDD building blocks and a state-of-play is given of the REDD negotiations, up to Durban. This focuses on the progress that has been made so far in crystallizing a workable REDD scheme. The REDD discussions focus on the five essential building blocks of any REDD scheme:

what kind of activities will be eligible and can be accounted for (scope);

will REDD be project-based (such as the CDM) or will it be al (sub-)national level (scale);

how will the efforts of developing countries be assessed and how will they be rewarded (accounting);

how will emission reductions be monitored and reported (MRV);

and where will the money come from to reward developing countries (finance).

Most developing countries do not have the technological and scientific capacity at this moment to monitor forests with sufficient accuracy at this stage. This could be a bottle neck for these developing countries to participate fully in a REDD scheme where payments will be results-based. Cheap but accurate forest monitoring tools are needed (Bucki et al., 2012). In this respect, the applicability of low resolution imagery in REDD is discussed.

Addressing the drivers of deforestation will be critical for national or sub-national governments of developing countries. In a globalised world, these have however become more elusive for national authorities. The potential undermining effect of increasing global demand for agricultural commodities on REDD is given. It will therefore be essential for the success of REDD that the financial and non-financial (e.g. ecosystem services of forests) reward for developing countries not only offsets the foregone benefits but also puts them on track of sustainable economic growth.

1. R EDD+ in the UNFCCC

1.1. Policy options for REDD+

Introduction

The design and details of the REDD+ decision, require careful consideration because it will have environmental, economic and social implications in the long term (Ghazoul et al., 2010). But REDD could also provide numerous co-benefits in addition to the reduction of greenhouse gas emissions (Stern, 2006). These include biodiversity and watershed conservation, renewable energy supply, food security for local communities and the maintenance of soil resources; and social benefits such as poverty alleviation and the protection of land and human rights (Dickson et al., 2009). These co- benefits will be achieved if the design of the international REDD+ framework provides incentives to project developers and governments across scales.

To date governments, non-governmental organizations (NGOs) and scientists have put forward many specific proposals on the design of a forestry-based mitigation mechanism (Parker et al., 2009). These proposals differ along a number of dimensions, such as the type of activities that will be included (scope), the geographical scale, the sources of funding and the approach of greenhouse gas accounting (CIFOR, 2010).

Many options have been proposed during the negotiation process and some have been placed in the background. For this analysis, we use a modular framework, drawn upon recent work undertaken by The Prince’s Rainforest Project and The Global Canopy Programme (Parker et al., 2009), consisting of five building blocks, representing the key constituents of any future international REDD mechanism e.g. scope, geographical scale, accounting mechanism, monitoring, reporting and verifying (MRV) scheme and financing mechanism (Figure 1). Using this framework as a guidance, the options for each building block that have emerged are compared.

The international design

of REDD+ will need to

provide incentives to all

developing countries,

bridging the diversity in

historical deforestation

rates and capacities to

implement forest policies

and monitor emissions.

SCALE

SCOPE

ACCOUNTING

MRV

FINANCE

Definition Activities Input - output Reference level Emissions - stocks

Monitoring Reporting

Source Distribution SAFEGUARDS

CDM

LULUCF

NAMA

Finance

Figure 1. Essential building blocks and elements of any REDD+ scheme. Links with other climate change discussions are outlined.

The scope of REDD Definitions

An important aspect is what is meant with forest (see also section 2.1). In total, there are over 890 different definitions of forests (Lund, 2005) and this has a significant impact on assessment of forest area (see Colson et al., 2009). The definition used in the UNFCCC is an area of more than 0.5–1.0 ha with a minimum “tree” crown cover of 10–30 %, with “tree” defined as a plant with the capability of growing to be more than 2–5 m tall. However, it also includes young stands of natural regeneration, all plantations which have yet to reach the required crown density or tree height, and areas normally forming part of the forest area which are temporarily unstocked as a result of human intervention (such as harvesting or natural causes) but which are expected to revert to forest. It would be logical that this definition (used in the Kyoto Protocol) will also be used in REDD+. However, this forest definition could lead to perverse incentives as it makes no distinction between natural forests rich in biodiversity and monoculture plantations. The inclusion of safeguards in REDD+, which include forest ecosystem services, has to some degree avoided the need of a more restrictive forest definition.

However, not only a clear definition of a forest is needed, but also for all other activities that are eligible under REDD+ (Sasaki and Putz, 2009; Simulu, 2009).

What kind of activities?

One of the key questions regarding REDD concerns the scope of land use and forestry activities that developing countries could undertake to contribute to climate change mitigation (Angelsen, 2008).

Since the beginning of the negotiation process in, 2005, there has been considerable interest in expanding this list of activities (Lawlor et al., 2010). The increasing number of activities included in the scope reflects the history of the policy debate, in which, at first, only avoiding deforestation (RED) was recognised as an important goal, to which avoiding forest degradation (REDD) was quickly appended. The additional elements making up REDD+, i.e. the conservation of forest carbon stocks, sustainable forest management and the enhancement of forest carbon stocks entered the debate at COP 13 in, 2007 (Herold and Skutsch, 2011). More recently, suggestions were made to expand the scope beyond forestry-based activities to also include CO2 and other greenhouse gas emissions from agricultural and other non-forest lands. This approach, which is based on a full land-based accounting of agriculture, forestry and land use (AFOLU), is termed ‘Reducing Emissions from All Land Uses (REALU) or REDD++ (Van Noordwijk et al., 2009). In the latter, all transitions in land cover that affect carbon storage are included, whereas in a REDD+ mechanism, developing countries can implement various activities that are voluntarily chosen from a list provided by the UNFCCC.

The first reason for expanding the list of activities was the recognition that forest degradation is a significant source of carbon emissions. Because the UNFCCC defines a forest as an area with a minimum crown cover of 10–30%, forests can be significantly degraded before they are considered deforested. Thus, without the inclusion of avoiding forest degradation as a goal, the integrity of the mechanism would be jeopardised. A second aim of broadening the scope was to minimise the potential for the displacement, or “leakage”, of emissions to land uses not included in the scope. The additional activities making up the ‘+’ would allow countries with varying rates of deforestation and levels of forest cover to participate, which will improve not only the international equity of the mechanism but also prevent shifting of deforestation activities to countries with low deforestation rates. Including agriculture, wetlands, peatlands and grasslands in the scope makes the mechanism even more effective at minimising leakage, because all displacements of emissions among sectors and land categories, including non-forest biomes, will be covered under AFOLU accounting (Lawlor et al., 2010). A third reason is the potential for maximising the climate contributions from the land-use sector by not only reducing negative changes but also by enhancing positive changes and, in the case of AFOLU accounting, additionally reducing emissions from agriculture (Lawlor et al., 2010).

However, broadening the scope also has some drawbacks. First, it increases complexity, particularly for measuring and monitoring, because four of the above-mentioned activities are not land-use change processes, as they constitute forest land that remains forest land, i.e. existing forests undergoing degradation, conservation, sustainable management and/or the enhancement of forest carbon stocks (GOFC-GOLD, 2009). Second, the UNFCCC does not have officially adopted definitions of forest degradation, sustainable forest management or the enhancement of forest carbon stocks (Griscom et al., 2009). The lack of universally agreed-upon definitions of these terms will cause complications when REDD is implemented on the ground (Sasaki and Putz, 2009). Third, uncertainty remains regarding the question of whether the term ‘enhancement of forest carbon stocks’ includes forest restoration only on lands already classified as forests or also includes the forestation of non- forest land (Angelsen et al., 2009b). A challenge associated with including afforestation/reforestation (A/R) is that it is already part of the CDM under the Kyoto Protocol. Excluding A/R, however, risks fragmentation of the overall forestry architecture (Angelsen, 2008). A last concern is that an expanded list of creditable activities can lead to the conversion of natural forests into plantations, because the UNFCCC definition does not recognise different biomes and, importantly, does not distinguish natural forests from plantations (Sasaki and Putz, 2009; Lawlor et al., 2010).

This definitional problem would be resolved in an AFOLU approach, because all transitions in land cover that affect carbon storage would be included (Van Noordwijk et al., 2009). Furthermore, REDD++ avoids fragmenting the framework into separate systems for different land-use categories.

However, the work leading to an integrated AFOLU framework will be complicated (Angelsen, 2008).

For this reason, including all carbon sources and sinks related to land use is supported as the ultimate objective of a future REDD mechanism but may be too complex to implement in the short term (UNFCCC, 2009).

The accounting mechanism specifies how emission reductions (ERs) and/or carbon stock enhancements are measured (Parker et al., 2009). The following sections show the steps and the options for designing such an accounting mechanism for REDD.

Accounting Basis

As a first step, the basis on which payments or credits would be made must be defined. Payments can be based either on the inputs needed to achieve a specific outcome or on the actual outcome (Angelsen, 2008). Input-based schemes suggest that financial benefits should be provided on the basis of developing countries’ implementation of policies and measures (PAM) to reduce emissions from deforestation and forest degradation, as in, for example, the ‘Compensated successful efforts’

proposal of Combes-Motel et al. (2009). Under an output- or performance-based approach, financial incentives are provided proportionally to the results achieved by REDD actions (Angelsen, 2008).

Payments would be granted for additional carbon sequestration or reduced emissions relative to a reference level calculated over a specified reference period (Santilli et al., 2005; Plantinga and Richards, 2008).

An advantage of the input-based approach is that deforestation levels need not be measured, whereas output-based approaches require a measurable output, ruling out certain countries that currently lack adequate technical capacity and institutions for these measurements (Alvarado and Wertz-Kanounnikoff, 2007). However, because financial benefits would be directly conditional on services delivered, such output can easily be connected to compliance markets, accommodating all sources of financing (see xxx), whereas input-based approaches can only be financed through a fund (Angelsen, 2008; Tacconi, 2009). Input-based programs would have some advantages in terms of lower transaction costs and the opening of separate negotiations regarding international forest sequestration and energy emissions (Aldy and Robert, 2008). An important limitation of the input- based approach is that it may be difficult to assess the likely impacts of many of the PAMs that countries might use to address the drivers of deforestation. Furthermore, reductions in greenhouse gas emissions could probably not be measured with sufficient certainty. The implementation of such PAMs should therefore be measured, reported, and verified using non-greenhouse gas metrics (Daviet, 2009).

Setting Reference Levels in an Output-Based Approach

A reference level can be defined as the level of a specified output (e.g. an emission or a carbon stock level), either at a point in time or over a period of time, that could be used to measure performance and/or award credits (Angelsen, 2008; Havemann, 2009). The reference level could be the same as a business-as-usual scenario (BAU), indicating a prediction of future emissions without REDD. However, proposals have been made to set the crediting reference level above (in the case of carbon stocks) or below (in the case of emissions) the BAU (Angelsen, 2008). Different proposals have been put forth for defining the key characteristics for setting reference levels, including the output and the reference period. These variables are explained in the next paragraphs.

Defining output

The output can be defined either as an emission (or a flow), as a carbon stock or as a combination of the two (Plantinga and Richards, 2008; Angelsen et al., 2009a). In the flow- or emission-based approach, only the net changes in carbon stocks for specific periods are used to calculate credits (Angelsen, 2008). In this case, the reference level indicates greenhouse gas emissions from deforestation and forest degradation in the absence of additional efforts to curb such emissions (Griscom et al., 2009). As an alternative or complement to the emission-based approach, suggestions have been made to connect incentives directly to the amount of forested area or forest carbon stock, regardless of a country’s past deforestation rates. The ‘stock-flow’ approach uses an reference level based on historical emissions but relies on a second instrument, i.e. a stabilisation fund, which provides payments for stocks but is not associated with credits. A fraction of the payments for reductions is withheld to raise funds to be distributed to forest countries in the form of payments for forest stocks (Cattaneo et al., 2010).

An emission-based approach was used in the Kyoto Protocol, making its application in REDD a natural step (Angelsen, 2008). However, opponents have pointed out the methodological challenges associated with emission-based approaches including data quality and availability, leakage control (Prior et al., 2007), and equity concerns such as accounting for early efforts in forest conservation (Cattaneo et al., 2010). Arguments in favour of the stock-based approach are the likely greater carbon effectiveness (Ashton, 2008) and the greater willingness of the private sector to pay (Prior et al., 2007). A risk of the stock-based approach is that payments could be made to forest areas that are not under threat, undermining the additionality of the mechanism (Angelsen, 2008).

Base Period

A reference level can be calculated based either on historical data or on projections of emission scenarios in the absence of a REDD mechanism. One of the earliest proposals for a mechanism to reduce emissions from deforestation, the ‘Compensated Reduction’ proposal of Santilli et al. (2005), suggested that reference levels should be based on a country’s average rate of emissions due to deforestation over a recent historical period. This idea is reflected in other national proposals, although the reference periods differ among them; e.g.

Santilli et al. (2005) suggested using a five-year average, whereas in other proposals, the reference period is ten years.

Historical deforestation rates are one of the best and easiest predictors of future deforestation in the short and medium terms (Angelsen et al., 2009b). They are also able to demonstrate “actual”

reductions relative to past emissions from deforestation (Parker et al., 2009). However, the historical approach also has some important limitations. First, historical reference levels will be difficult to establish with accuracy in countries with limited comparable (across years) and/or unreliable high- resolution historical deforestation data (CIFOR, 2010). Second, because deforestation dynamics and the timing of deforestation differ greatly among countries and even within countries, the reference period chosen to estimate an reference level will make a great difference (Corbera et al., 2010).

Regarding overall equity, this methodology is likely to provide incentives only for countries with high historical rates of deforestation (Alvarado and Wertz-Kanounnikoff, 2007). In addition, when positive incentives are extended only to these historically high-deforestation countries, there could be an exacerbated threat of the leakage of deforestation activities to countries with historically low

Developing countries are in different

phases on the forest transition curve, a

single method for setting the reference

level and assessing emission reductions

therefore cannot provide incentives to all.

deforestation rates, including “high-forest, low-deforestation” (HFLD) countries (Cattaneo et al., 2010). Finally, this approach does not recognise potential changes in national circumstances over time such as changes in the rates deforestation and economic growth (Parker et al., 2009). In particular, if there is evidence that deforestation is likely to decline in any of the large remaining tropical forest areas, a major risk is the creation of emission allowances that are not additional, i.e.

that are “hot air”, which undermines the environmental integrity and credibility of REDD (CIFOR, 2010; Corbera et al., 2010).

Several proposals have attempted to address the limitations of the historical approach by the use of an adjustment factor for the historical deforestation rate in establishing reference levels. To address differences among countries, incorporate potential changes in national circumstances and address the potential for leakage, two main options have been proposed. The first is extending reference levels that are higher than the historical values to countries with historically low deforestation rates, and the second suggestion is to apply a development adjustment factor to the historical reference level (Parker et al., 2009).

In the first approach, countries with high rates of deforestation would be rewarded for reducing these emissions under a crediting reference level set relative to their historical deforestation rates. In contrast, countries with low rates of deforestation would gain credits for preventing emissions growth relative to reference levels that are elevated based on assumptions of increased future rates of deforestation (Cattaneo et al., 2010). One option is to use the reference emission rate indexed to the global deforestation rate for countries with little or no historical deforestation, as suggested in the ‘incentive accounting’ proposal (Mollicone et al., 2007). Countries with deforestation rates of less than half the global average would use the historical global deforestation rate as their national RL, whereas countries with higher deforestation rates would use their national historical reference level (Mollicone et al., 2007). As an alternative, Strassburg et al. (2009) proposed a ‘combined incentives’

mechanism, in which incentives allocated to an individual country are determined by a formula that combines a measure of individual country performance against their own historical reference level, and performance against a global RL. In this approach, different scenarios are generated by differing the weights put on historical global deforestation and national deforestation for different types of countries (Strassburg et al., 2009).

In the second option a development adjustment factor is applied to the historical reference level to reflect predicted changes in future drivers of deforestation and which takes into account national circumstances such as forest cover, income (Gross Domestic Product) per capita, demographic trends, agriculture and infrastructure development (UNFCCC, 2007; Parker et al., 2009). Adding a growth cap to the reference scenario allows for certain amounts of deforestation to occur for the purpose of a country’s socio-economic development (Angelsen, 2008).

A general advantage of these measures is the upward adjustment of the national reference level for countries with lower-than-average historic deforestation rates as an incentive for maintaining these low rates, which will make it more attractive to these countries to participate in REDD (Murray, 2008). By including countries at all stages of the deforestation process, it will minimise the threat of leakage and increase the overall effectiveness and equity of the mechanism (Alvarado and Wertz- Kanounnikoff, 2007; Strassburg et al., 2009). Nevertheless, the risk of “hot air” remains with the use of an adjustment factor to the historical reference level. When setting global deforestation rates to define national reference levels, many countries are likely to receive reference levels above their emissions in a BAU scenario, eventually giving a sum of national reference levels higher than the global reference level. This increases the possibility of generating more credits than emission reductions at the global level, compromising additionality (Angelsen, 2008; Mur,ray, 2008; Busch et al., 2009). The ‘combined incentives’ mechanism is the only option that addresses this problem by

setting the sum of national reference levels equal to the global reference level through a flexible combination of global and national reference level determination.

A projected reference level reflects future emissions from deforestation, usually based on past deforestation, and projections for key social, economic, political and technological variables (Eliasch, 2008). They aim is to predict how deforestation rates are most likely to change in the future, employing simulations of change in land use and land cover (LUCC) models (Alvarado and Wertz- Kanounnikoff, 2007). Ashton et al. (2008) have proposed a ‘forward-looking’ reference level that uses the fraction of the volume of terrestrial carbon stock estimated to be at risk for emission in the long run, based on biophysical, economic and legal considerations.

If it were possible to predict when and where deforestation would occur without an incentive scheme, additionality would be maximised (Eliasch, 2008). However, LUCC models carry the risk of inaccuracy given the uncertainties regarding the evolution of direct and indirect causes of deforestation, including the evolution of agricultural commodity prices, biofuel markets and the production of timber goods (Alvarado and Wertz-Kanounnikoff, 2007). Conversely, their ability to incorporate country-specific circumstances such as drivers of deforestation makes this approach very attractive to countries in different stages of the deforestation process, minimising the risk of leakage (Huettner et al., 2009). Still, calculating reference levels based on future projections demands sufficient technical capabilities, and the complexity of LUCC models limits their transparency and clarity to policy makers (Huettner et al., 2009).

Monitoring, Reporting and Verification Scheme Monitoring

To participate in REDD, countries will need to establish a (national) MRV system to estimate anthropogenic forest-related greenhouse gas emissions from sources and removals by sinks. Such a system should support the MRV requirements under the UNFCCC, so that such estimations are based on a common international methodological approach for MRV in REDD (Maniatis and Mollicone, 2010). Until now, the proposed options for estimating forest-related greenhouse gas emissions within a REDD mechanism are restricted to the most recent Good Practice Guidance (GPG) (2003) and Guidelines (2006) of the Intergovernmental Panel on Climate Change (IPCC), although other carbon-estimating methods exist, e.g. the CDM methodologies and the Voluntary Carbon Standard methodologies (Bird et al., 2010). Because the latter two have not been submitted to the UNFCCC, only the application of the IPCC GPG and Guidelines are assessed here.

The IPCC GPG and Guidelines refer to two basic inputs used to estimate greenhouse gas inventories:

activity data and emission factors. To represent activity data or changes in areas of different land categories, three different Approaches are defined. These range from the collection of non-spatial country statistics to spatially explicit land-conversion information, derived from sampling or wall-to- wall mapping techniques in the third Approach (Angelsen et al., 2009b). Secondly, the emission factors refer to the emissions or removals of greenhouse gases per unit of activity and can be measured as changes in carbon stocks in the various carbon pools of a forest, e.g. aboveground biomass, belowground biomass, litter, dead wood and soil organic carbon. There are three Tiers of data for emission factors in the IPCC GPG and Guidelines, representing increasing levels of analytical complexity and data requirements, ranging from default values to actual inventories, using repeated measures of permanent plots to directly measure changes in forest biomass and/or well- parameterised models in combination with plot data (IPCC, 2006).

The benefits of using these methods as the basis for REDD are that they have been developed and reviewed by experts, they have already been accepted in UNFCCC negotiations and they have been

tested during the initial phase of the Kyoto Protocol (Olander et al., 2008). However, monitoring for REDD would require identifiable and traceable land conversion, and only Approach 3 will accommodate this (Baker et al., 2010). Furthermore, although moving from Tier 1 to Tier 3 increases the accuracy and precision of the greenhouse gas estimates, it also increases the complexity and the costs of monitoring (Maniatis and Mollicone, 2010). According to DeFries et al. (2007), the capacity for deforestation monitoring is well advanced in a few developing countries and is a feasible goal for most others. Nevertheless, an assessment of current monitoring capabilities conducted by Herold (2009) emphasised that the majority of countries have limitations in their abilities to provide a complete and accurate estimation of greenhouse gas emissions and forest loss. Consequently, given the differences in ecology, institutions and technical capabilities among countries, no single MRV system will apply to all developing countries (Baker et al., 2010). A tiered system would provide some flexibility for differences in technical capability among countries (Olander et al., 2008). However, Tier 1 estimates, based on default values, can have uncertainties as large as 70% (Angelsen et al., 2009b).

Clearly, substantial improvements (at least to Tier 2) over that value will be required if the participating countries are to meet international compliance standards (Baker et al., 2010).

Financing Scheme

Generating financing for REDD activities at an adequate and sustainable scale is crucial for creating incentives and payment systems for government actions and specific projects to reduce emissions that overcome the drivers of deforestation and forest degradation (Evidente et al., 2009). The Eliasch Review (2008) on financing global forests estimated that US$ 17–33 billion must be invested annually to halve greenhouse gas emissions from deforestation by 2030. Once generated, this financing must be distributed among the participating countries. The generation and the distribution of payments can be seen as two more-or-less independent mechanisms. For this reason, the two mechanisms will be addressed as two sub-building blocks of the overall financing scheme.

Finance Sources

The source of finance refers to the type of economic instrument that the mechanism uses to generate revenue (Parker et al., 2009). Recognising the varied interests and institutional capacities of countries, various proposals are being discussed as potential financing sources (Evidente et al., 2009).

These sources can be classified into three main instruments: fund-based instruments, market-based instruments and combinations of the two instruments in hybrid approaches.

In the fund-based system, incentives for REDD would be paid by a fund, which could be made up of voluntary financial contributions or provided through market-linked instruments. The emission reductions generated in this approach cannot be purchased as offsets by developed countries to meet their national targets (Skutsch and McCall, 2010).

A first fund-based option for providing REDD financing is for governments, financial institutions or private entities to voluntarily contribute to a fund, which can then be distributed to participating developing countries to aid and reward their efforts to reduce emissions from deforestation and forest degradation (Evidente et al., 2009). The manner in which these contributions are resourced can take a variety of forms, ranging from, e.g. bilateral or multilateral commitments to a global mechanism (Daviet et al., 2007; Evidente et al., 2009). A voluntary fund-based approach offers three advantages: first, it allows for differentiation among types of forests, policies and programs by taking into account each country’s particular circumstances; second, it decouples carbon accounting and quantification issues from Annex I parties’ emission-reduction targets; and third, it provides the required up-front payments for capacity building and the implementation of REDD activities (Daviet et al., 2007). The major arguments against a solely fund-based mechanism are that it is unlikely to be able to generate funding at the required scale to effectively provide support for emissions reductions

activities and that funding will be difficult to continue over the long term (Evidente et al., 2009;

Corbera et al., 2010). For this reasons, market-linked instruments have been proposed. These instruments raise revenue indirectly from the (carbon) market through a variety of mechanisms (Parker et al., 2009). For example, proposals have been made to tax carbon-intensive commodities and services in Annex I countries and to set a levy on the sale of Assigned Amount Units (AAUs) or on transactions involving Emission Reduction Units (ERUs); this revenue would be placed into a carbon fund (Corbera et al., 2010). Such a levy on an auction process, either at the international level, as proposed in Norway, or at the national level, as in Germany’s “International Climate Initiative”, could generate revenues at the necessary scale, and would ensure additional emission reductions to existing commitments (Parker et al., 2009). However, these market-linked instruments would depend on the existence of a sound long-term (carbon) market to produce a predictable flow of funds and could be politically difficult to negotiate (Corbera et al., 2010).

Market-based approaches generally refer to a mechanism whereby participating developing countries are able to create and sell emission-reduction credits. There are three types of market- based instruments: compliance-based markets, separate markets and voluntary carbon markets.

The first market instrument creates incentives that are linked to compliance with developed countries’ commitments to greenhouse gas emission reductions. In this option, the forest emission- reduction credits sold by developing countries would be fully fungible in the global carbon market and used for compliance by either governments that have binding targets under a post-2012 climate change agreement or by companies in trading systems such as the EU Emission Trading System (ETS) (European Commission, 2008). According to several authors, including Corbera et al., (2010) and Pedroni et al., (2009), market instruments are the most significant sources of funding, because carbon offset markets have the potential to assure long-term, continuous and predictable flows of finance for REDD. However, concerns exist that a direct integration of REDD into carbon markets may result in a destabilisation of these structures. The introduction of potentially large volumes of low- cost credits could create market disruptions and increase price volatility (Ogonowski et al., 2007;

Angelsen, 2008). It would also weaken the incentives for abatement in developed countries and could result in a significant delay of mitigation actions by the most important emitters (Wertz- Kanounnikoff and Tubiana, 2007; CIFOR, 2010).

Various options have been put forward to reduce the risks of integrating REDD into carbon markets including the adoption of deeper emission-reduction commitments by developed countries, limiting the supply of REDD credits into the market and controlling the interchange ability of REDD credits in the form of separate but linked markets or the creation of a new trading unit (Angelsen, 2008; CIFOR, 2010). In a separate market, REDD credits are linked to existing emission-reduction credits in other countries and sectors but have varying degrees of fungibility (i.e. tradability) (Parker et al., 2009). Examples of separate markets for REDD are the ‘Dual-markets approach’ (Ogonowski et al., 2007) and the ‘Tropical Deforestation Emission-Reduction Mechanism (TDERM)’ proposal (Hare and Macey, 2007). The latter involves the creation of a special trading unit, the Tropical Deforestation Emission Reduction Unit (TDERU), which would represent both emission reductions and other ecosystem services. Developed countries would

A market-based REDD+ scheme could

generate the flow of money needed to

make REDD+ a success but it will need

to be imbedded with great care in an

overarching carbon market to avoid

that low-cost REDD+ credits delay

mitigation actions in developed

countries and in other sectors.

take on commitments to purchase TDERUs as part of their overall emission-reduction commitments (Hare and Macey, 2007).

A third market mechanism, for which the procedures are not as lengthy, costly or binding, consists of voluntary carbon offset schemes. These make it possible to sell carbon credits to individuals, companies or entities (large towns, institutions, etc.) wishing to compensate for emissions linked to their activities (Karsenty, 2009). Although the voluntary carbon markets are playing a pioneering role in market-based initiatives to foster investment in REDD, they are unlikely to mobilise sufficient funding to finance widespread REDD adoption (Angelsen et al., 2009b).

General concerns about market-based approaches are related to the required upfront financing for capacity building and separate financing necessary to address the broader social and political factors that contribute to deforestation such as land tenure and indigenous peoples’ rights, enforcement and monitoring capabilities and coherent economic and agricultural policies (Peskett et al., 2008;

CIFOR, 2010). Further concerns are related to fears that a market is incompatible with forest conservation and that it will have a negative impact on ecological and other forest values. At a higher level, there is concern that in a market mechanism, the more powerful players may benefit most, both within and among countries, as has been the case in CDM. Countries which have relatively good forest inventory data and sufficient technical capacity, e.g. Brazil and Mexico, will be able to take advantage of a REDD market quickly, whereas others may fall behind (Skutsch and McCall, 2010).

Noting that both market and non-market systems have their limitations, various combinations have been explored in attempts to use the strengths of each system (Evidente et al., 2009). As a first hybrid funding option, the type of financing could depend on the type of action being undertaken. In this dual system, contributions from government funds would finance government activities—

improving forest policies and governance, land-tenure reform and indigenous rights, agricultural and economic policies, among others—while market-linked or market financing would direct resources to people and communities to provide incentives and to support activities that directly result in emission reductions at the ground level (Viana, 2009).

A second hybrid funding option that has gained significant support during the negotiation process is the ‘phased approach’. Due to the different levels of institutional capacity and development in the developing countries, a REDD framework should be implemented in three phases. The First Phase, preparation and readiness, includes the development of national strategies or action plans, policies and measures and capacity building. This is followed by a Second Phase that focuses on the implementation of PAMs, addressing the drivers of deforestation and demonstration activities for emission reductions. The Third, and last, Phase includes the full implementation of a greenhouse gas -based instrument that rewards performance on the basis of quantified forest emissions and removals with respect to an agreed-upon reference level (UNFCCC, 2009). In this scenario, the sources of financing vary according to the phase of REDD implementation, starting with activity- based payments, mainly provided through non-market based funds, with a transition to result-based payments in the Second Phase. In the latter, financing may come from either funds or market sources. An option to finance the intermediate phase is the creation of a separate market for REDD, because this will still provide incentives for the private sector to invest while allowing a REDD market to stabilise before any full-scale linking with the post-2012 global carbon market (Ogonowski et al., 2007). When institutions develop sufficient capacity for monitoring and demonstrating emissions reductions, countries could proceed to a full implementation phase, in which payments will be entirely based on results and financing by compliance markets becomes feasible (Angelsen et al., 2009b; UNFCCC, 2009). The advantage of the phased approach to REDD lies in its flexibility (Angelsen et al., 2009b); countries can participate according to their capacity and have incentives to progress from one stage to the next. This means that a wide range of tropical-forest countries will be able to take part in REDD. Countries with sophisticated MRV systems and sound institutional frameworks

may start at Phase Three. Other countries with less sophisticated MRV systems can start at Phase One or two but have incentives to move towards more sophisticated systems so that they can graduate to Phase Three. The incentive for graduating from phase one to phase three is that by doing so, countries generate a more reliable income from REDD (Angelsen et al., 2009a).

Distribution System

The second sub-building block of the financing scheme explores the different options for delivering climate financing to developing countries and the equality of the distribution of REDD revenues across countries. In general, equity concerns are addressed implicitly in the reference level methodology (Parker et al., 2009). Examples are adding a development factor to the reference level or the ‘combined incentives’ and the ‘incentives accounting’ approaches. There are, however, also mechanisms to redistribute revenues independent of the methodology used to set the reference level. One mechanism withholds a proportion of the revenues for emission reductions (as a levy or tax), which feed into a fund for REDD countries in the form of stock payments (Cattaneo, 2009). In both of these approaches, the revenue required to support HFLD countries is generated from the mechanism itself (Parker et al., 2009). Potential disadvantages of these approaches are the distorting effect that redistribution could have on incentives to reduce emissions in countries with high rates of deforestation (Parker et al., 2009). As an alternative, an additional distribution mechanism or

“stabilisation fund” was proposed, which would use additional funding, generated outside the REDD mechanism, to address leakage and equity concerns in HFLD countries. The revenue for a stabilisation fund could come from a variety of sources including voluntary funds or innovative financing mechanisms such as the auctioning of allowances or levies on shipping or aviation (Parker et al., 2009).

Scale

The scale of REDD refers to the geographic scale at which eligible REDD activities will be implemented, MRV will be performed and an international funding mechanism will provide incentives for REDD activities. Four main options for the scale of REDD are proposed: the project level, the sub-national level, the national level and an integrated approach (Angelsen, 2008; Cortez et al., 2010).

Several countries, such as the United States and Colombia, have firmly expressed being in favour of a project-based approach, similar to that of the CDM (Calmel et al., 2010). In this approach, incentives flow directly to project developers based on performance against a project reference level. Although each project will not necessarily coincide with a governmental jurisdiction, a sub-national implementation of REDD nevertheless indicates that incentives will flow to a sub-national government entity such as a state, municipality, province or district based on their performance compared with a sub-national reference level (Cortez et al., 2010). In the national approach, a country would be required to establish a national strategy and will receive payment if their emissions are reduced in comparison with the national reference level. A national emissions monitoring system would also be established to verify that these emission reductions are additional (Calmel et al., 2010). The principle of the national approach is that the beneficiary of the carbon credits generated is the country, which would be responsible for distributing them among the concerned stakeholders in accordance with procedures established during the preparation phase (Calmel et al., 2010). An

‘integrated approach’ to REDD has been proposed as an option for creating incentives for action at multiple scales (Cortez et al., 2010). Under an integrated approach, sub-national activities are integrated into a national accounting framework (Angelsen, 2008). Both projects and/or sub-national and national activities can be started immediately (Pedroni et al., 2009). The national government would set up a national accounting framework and establish a nationwide monitoring system.

Simultaneously, implementation of REDD+ activities would also occur at the sub‐national level, led by

local/regional governments, communities, NGOs or private developers (Cortez et al., 2010). Different versions of an integrated approach have been presented. In the ‘transitional approach’, the sub- national level is seen as an intermediate phase. Developing countries would be able to decide on their initial level of participation in this mechanism according to their particular circumstances and interests (Pedroni et al., 2009). However, in the case of an implementation of activities at the sub- national level, a country would need to scale up to a national approach as they strengthen their capacity and improve governance. Transition to a national approach would be obligatory, either within an agreed time frame or when an agreed percentage of forest area is covered by REDD activities (Angelsen, 2008). In the ‘nested approach’, the sub-national level continues to exist and can still be credited after a national accounting framework has been established (Cortez et al., 2010).

The decision on the scale of the REDD mechanism involves a trade-off between the capacity of a developing country to participate in the scheme, which may depend on, e.g. data availability, institutional and financial capabilities and the potential risks of leakage. Implementation at a lower scale, i.e. projects and sub-national activities, would initially be easier than national approaches and would accommodate different national circumstances and account for intra-national heterogeneity in the capacity to implement REDD projects (Myers, 2008). Conversely, opponents of these lower- scale approaches note that they face greater challenges in addressing leakage and permanence than do national-level approaches (Cortez et al., 2010). However, the national-based approach does not entirely solve the leakage problem, because the issue of international leakage remains as emissions could shift from participating to non-participating countries (Corbera et al., 2010).

A second challenge is to find a balance between triggering the policy reforms required to address land-use change drivers and to involve a broad scope of stakeholders and so sustain a long-term success. Because the causes of land-use change are many and variable and, in some cases, are even linked to national-level policies, the involvement of national governments is essential to achieve the necessary large‐scale systemic policy reforms (Eliasch, 2008; Virgilio et al., 2010). Furthermore, given the magnitude of emissions, the implementation of REDD at only low levels risks having a small impact and is not likely to address the broader drivers (Myers, 2008; Virgilio et al., 2010). Conversely, according to Cortez et al., (2010), a greater participation by actors with direct control over land‐use decisions, including sub‐national governments, indigenous peoples and forest‐dependent communities, and landowners/users, can be motivated by providing direct incentives for sub‐national activities. Nevertheless, the national approach addresses sovereignty issues by creating country ownership, and it would give governments the flexibility to establish a broad set of PAMs to reduce deforestation and forest degradation (Angelsen, 2008). In addition, a national approach can be aligned with national development strategies, possibly bringing long-term development benefits (Angelsen, 2008; Eliasch, 2008).

A third issue in the scale debate is the question of cost efficiencies versus the generation of the significant up-front capital needed to enact REDD programs. While a national approach enjoys significant economies of scale such as lower transaction costs and lower MRV costs, many potential private-sector investors are hesitant to invest up‐front capital in national programs because of concerns over controlling risks, lack of transparency, poor governance and corruption, among others (Cortez et al., 2010).

Regarding social equity, because incentives must reach local actors in the deforestation process, the implementation of national approaches as an exclusive instrument to provide incentives to reduce emissions from deforestation in developing countries could have negative intra-national equity implications (Eliasch, 2008; Corbera et al., 2010). Although some developing countries may have transparent systems for benefit sharing already in place, others, however, lack the institutional capacity and legal safeguards to ensure that a centralised REDD scheme would equitably allocate resources to local actors (Costenbader, 2010).

An integrated approach to REDD has the potential to address many of the drawbacks of pure national or sub-national approaches by accounting for in-country leakage, engaging national governments, and taking advantage of certain economies of scale, while also motivating sub-national actors to participate in REDD and attracting greater private investment. An integrated approach may also provide for a more transparent distribution of the benefits from REDD, because local actors could own and transact credits directly rather than relying on a national system of benefit sharing (Cortez et al., 2010). Further, the sub-national activities provide important learning opportunities for countries to test options for building national capacity and institutions regardless of whether or not the sub-national level would be an intermediate phase (Virgilio et al., 2010). However, a nested approach may be more difficult from an institutional point of view, because the sub-national and national accounting systems would need to be harmonised, and a framework for transferring REDD incentives across scales must be defined (Angelsen, 2008; Corbera et al., 2010). Furthermore, risk- management mechanisms would need to be developed to mitigate the risk of revenue loss by sub‐national entities in the case of national-scale non‐performance (Cortez et al., 2010).

Conclusion

The negotiation process for an international REDD mechanism has so far resulted in a series of proposals on policy approaches and positive incentive mechanisms. The latest decision document for a REDD+ mechanism, adopted at COP 16 in Cancun, Mexico (UNFCCC, 2011), represents a key step in achieving an overall land use and forestry-based mitigation framework.

Although some decisions were made about REDD in Cancun, e.g. the scope of REDD and its scale of implementation, the negotiation of a forestry-based mitigation mechanism will be a continuous process. This paper is intended to inform this ongoing negotiation process regarding the inclusion of REDD within a future climate change agreement.

By dividing the REDD mechanism into five modular building blocks, an attempt was made to provide a complete overview of the proposed options for each building block, including those that have been placed in the background of the current negotiations. The assessment herein shows that these options face several design and implementation challenges in terms of mitigation potential, abatement costs, environmental risks and benefits, social equity and institutional feasibility. As a result, when assembling the overall REDD framework, all these implications will involve trade-offs to enable the design of a sustainable international REDD framework that will deliver effective, efficient and equitable results and that will be feasible for all developing countries.

1.2. REDD+ negotiations: a state-of-play

From Montreal to Copenhagen

At the 11^th Conference of the Parties (COP11), Papua New Guinea and Costa Rica submitted on behalf of the Coalition for Rainforest Nations (CfRN) a proposal to establish a mechanism called ‘Reducing Emissions from Deforestation in developing countries (RED)’ under the UNFCCC (Lawlor et al., 2009).

The CfRN was established several months before the COP, and unites countries aiming to promote a more sustainable use of tropical forests. The aim of the RED mechanism would be to curb or limit deforestation and its related greenhouse gas emissions, by providing economic incentives to developing countries to keep their forests intact (Karsenty, 2008). Discussions were initially limited to reducing emissions from deforestation, but expanded quickly to include forest degradation (REDD) (Verchot and Petkova, 2009).

4/2008

Norway’s Climate and Forest Initiative 6/2008

FCPF

2004 2012

2005 Montreal

2007 Bali

2010 Cancun

2011 Durban 2008

Poznan

2009 Copenhagen

5/2010 REDD+

partnership 5/2005

CfRN

9/2008 UN-REDD

Figure 2. Timeline of the most important UNFCCC meetings and initiatives with respect to REDD+ [CfRN Coalition for Rainforest Nations, FCPF Forest Carbon Partnership Facility].

The principle of REDD was accepted at COP13 of the UNFCCC held in December, 2007 in Bali. The Parties agreed in the Bali Action Plan (BAP) that a REDD scheme should be one of the building blocks of a new climate agreement (as part of the Ad-Hoc Working Group on Long-term Cooperative Action). Further, the BAP encourages Parties to expand the scope beyond deforestation and forest degradation and to explore options to include forest conservation, sustainable management of forests and the enhancement of forest carbon stocks (UNFCCC, 2007). These new activities were expressed as the “+”, and since then, 2008 COP14 in Poznan REDD+ is officially defined as ‘reducing emissions from deforestation and forest degradation in developing countries, and the role of conservation, sustainable management of forests and enhancement of forest carbon stocks in developing countries’.

In the BAP, the Parties to the UNFCCC decided to launch a negotiation process for a post-2012 climate change agreement scheduled to be concluded at the COP15 in December, 2009. This agreement should also include financial incentives for forest-based climate change mitigation actions in developing countries. At COP15 in Copenhagen in, 2009 however, countries could not find a common ground for an overall legally-binding climate change agreement only for a general Copenhagen Accord. REDD+ was one of the few issues though on which significant progress was